Short stature in competitive prepubertal and early pubertal male gymnasts: the result of selection bias or intense training?

OBJECTIVES: To determine whether high-volume, high-impact physical training in prepubertal and early pubertal male gymnasts is associated with reduced statural and segmental growth and reduced serum insulin-like growth factor-I (IGF-I) and increased cortisol (C) levels. STUDY DESIGN: Height, sitting height, leg length, and segmental lengths (humerus, radius, femur, and tibia) and breadths (biacromial and bi-iliac), diet, serum IGF-I, testosterone, and C were measured in competitive male gymnasts and normoactive children (Tanner stage < or = 2) every 3 to 4 months over an 18-month period. RESULTS: At baseline, gymnasts (n = 31) were 0.7 years older than members of the control group (P <.05, n = 50) but were no different in terms of biologic maturity. Age-adjusted z scores showed that the gymnasts were shorter than members of the control group (-0.5 +/- 0.2 SD, P <.05) because of reduced leg length (-0.8 +/- 0.2 SD, P <.001) but not sitting height. Segmental lengths and bi-iliac breadth age-adjusted z scores were also reduced in the gymnasts (P ranging <.05 to <.001). No difference was detected for serum IGF-I or C. After 18 months of follow-up, no differences were found for rates of change in height, sitting height or leg length, segmental lengths, IGF-I, or C between those gymnasts and control subjects who remained prepubertal and early pubertal (gymnasts n = 18; control group n = 35). However, the magnitudes of baseline differences in anthropometric measures (z scores) persisted throughout the study. CONCLUSION: Short stature in these competitive male gymnasts was due to a reduced leg length but not sitting height. The lack of a difference in growth rates, IGF-I, and diet over the 18-month period indicates that the short stature reported in male gymnasts is due to selection bias rather than gymnastics training.